The questioner didn't specify whether the water will be cleaned up in a "batch" for a single household; or, whether treatment would be done continuously, for a piped water distribution system. We'll suppose the batch approach.
Mechanical, single-stage filters, employing commonly available materials such as sand, activated charcoal, clay, peat, straw, or other agricultural waste, can be designed to effectively remove large particles and some larger bacteria and parasitic organisms. After an extended period of use, the pore structure of a mechanical filter tightens up, and commensurately removes some smaller particles; but, this characteristic is not suitable for removal of all viruses, smaller parasitic organisms, and very fine solids. Unless you can afford more high tech approaches such as polymeric membrane filtration, ceramic, or molecular sieve filter media, which all have very limited flow-through capacities and which are expensive and consume electricity, filtration alone will not likely meet all the needs you've identified.
With the possible exception of granular activated carbon filter media --- (unfortunately, GAC has the drawback of being produced through an energy intensive and highly C02 emitting manufacturing process) --- no purely mechanical batch filter is capable of reliably removing dissolved heavy metals over an extended period of use.
Cautions: The ratio of dissolved to particulate forms of heavy metals in well water may vary over time, depending on rainfall/recharge and well withdrawal rates. In a batch filter there is a risk that dissolution of a particulate form of a heavy metal may take place within the filter media unless the design is based on a thorough understanding of local water chemistry and what bacteria are doing inside the filter.
Removal of heavy metal salts require chemical treatment or ion exchange. This could be done with a single packet of chelating/precipitation additive put in prior to filtration, or an ion exchanging material introduced into the filter media (as in the bed of nails example in the CSM story linked to above), or a second stage chemical precipitation and settling step, for examples.
Arsenic is seldom found in surface water in health threatening concentrations unless there is pollution source: a pesticide spill, a leaching mine waste heap, a chemical plant, or whatever. When well water has dangerous levels of arsenic it may be from excessive withdrawals that are raising deep naturally arsenic containing waters toward the well point. (Called saline intrusion.) This is what happened extensively in parts of Asia (see CSM article for background). If heavy metals can be abated through upstream pollution controls or through modifications to local water well designs and use configurations, chemical treatment may not be needed. Assuming you find that chemical treatment is needed, here are two more places to look for guidance.
First we suggest you find out what treatment/filtration technologies have been found most affordable and reliable in Nepal, India, or Bangladesh. Being developing countries with a large scale arsenic contamination problems to overcome, their examples are likely your best options.
Another suggestion we have is to contact your State public health agency and/or USEPA to find a copy of the cost effectiveness studies EPA performed in promulgating drinking water standards (maximum contaminant levels) for arsenic. This would be applicable in high density, more urban situations. What you'll want is the table that identifies annualized cost per thousand gallons treated for the best available technologies. Note that there may some controversy as to whether the appropriate maximum concentration should be the one EPA chose as an end point to protect health or, rather, the one that the World Health Organization chose. You can probably guess, without even looking, which entity proposes a lower threshold number for protecting human health.
When choosing a target maximum drinking water concentration for arsenic, for example, think about what cultural assumptions went into the risk assessments that underlie them. The US standard, for example, may have been set based on assumptions about how much water is absorbed in a typical American family boiling their food and how much bottled beverages are consumed on average. These assumptions may have little applicability in rural Nicaragua. Just be sure to think it through.